MODEL REFERENCE / tencent

tencent/Hy3-preview

Tencent Hunyuan Hy3-preview — scaled-up MoE language model (295B total / 21B active) with a 3.8B MTP layer for speculative decoding, 256K context, and hy_v3 tool/reasoning parsers

Hunyuan Hy3-preview MoE — 295B/21B on 8×H200, 8×H20-3e(141GB), or 8×AMD MI300X/MI355X with MTP

View on HuggingFace View on ModelScope

vLLMmoe295B / 21B262,144 ctxvLLM 0.20.0+text

02 / Configure

Configure serving

Choose engine, hardware and deployment options. The command updates as you go.

03 / Understand

Technical notes

Model notes and examples